Have been on a ver9 NE with NG backup module in use and upgraded to 10.1.8 before EOL end of June. Setup the NE backup (standard) and have been taking the normal weekly full and nightly incrementals.
But from reading the administrator guide and looking at some of the wiki articles on the new (old) backup system and how it handles incrementals I'm a bit concerned we're falling through the cracks with redo logs. The default value for zimbraRedoLogRolloverFileSizeKB is 1GB and our daily email volume is not enough to reach that size, have it be rotated and archived. The docs seem to be saying that incremental backups will copy any archived redo logs into the backup target folder, but nothing talks about what happens to that main redo.log in terms of backup - it seems to just get removed/restarted.
Does that mean that when the nightly incremental job kicks off:
We don't have an /opt/zimbra/backup/redologs/ folder and none of the incremental folders under the sessions subfolder seem to have anything recognizable as a redo log stored.
Should we be reducing the zimbraRedoLogRolloverFileSizeKB back to its old 100MB value so we have archives to be moved by the incremental process?
Given we're a fairly small system (<100 mailboxes) I'm wondering if I just do nightly full backups and skip incrementals.
I'll spin up a test install to test a restore of incrementals shortly (I'm also snap-shotting the VM nightly so I've got another backup method while this is bedded down).
Here's the zmprov config
And here's an incremental success email with the weird redo log sequence range that first caught my attention and made me look a bit closer. The full backups seem to have proper numbers in the range (in the 500s).
But from reading the administrator guide and looking at some of the wiki articles on the new (old) backup system and how it handles incrementals I'm a bit concerned we're falling through the cracks with redo logs. The default value for zimbraRedoLogRolloverFileSizeKB is 1GB and our daily email volume is not enough to reach that size, have it be rotated and archived. The docs seem to be saying that incremental backups will copy any archived redo logs into the backup target folder, but nothing talks about what happens to that main redo.log in terms of backup - it seems to just get removed/restarted.
Does that mean that when the nightly incremental job kicks off:
- there are no archived redo logs, so nothing is moved into the backup target
- and it looks to restart the main redo.log file
We don't have an /opt/zimbra/backup/redologs/ folder and none of the incremental folders under the sessions subfolder seem to have anything recognizable as a redo log stored.
Should we be reducing the zimbraRedoLogRolloverFileSizeKB back to its old 100MB value so we have archives to be moved by the incremental process?
Given we're a fairly small system (<100 mailboxes) I'm wondering if I just do nightly full backups and skip incrementals.
I'll spin up a test install to test a restore of incrementals shortly (I'm also snap-shotting the VM nightly so I've got another backup method while this is bedded down).
Here's the zmprov config
Code:
zimbraRedoLogArchiveDir: redolog/archivezimbraRedoLogCrashRecoveryLookbackSec: 10zimbraRedoLogDeleteOnRollover: FALSEzimbraRedoLogEnabled: TRUEzimbraRedoLogFsyncIntervalMS: 10zimbraRedoLogLogPath: redolog/redo.logzimbraRedoLogRolloverFileSizeKB: 1048576zimbraRedoLogRolloverHardMaxFileSizeKB: 4194304zimbraRedoLogRolloverMinFileAge: 60Code:
Label: incr-20250603.040626.405Type: incrementalStatus: completedStarted: Tue, 2025/06/03 14:06:26.405 AESTEnded: Tue, 2025/06/03 14:06:30.513 AESTRedo log sequence range: -1 .. ?Number of accounts: 73Statistics: Posted by pixelplumber — Tue Jun 03, 2025 12:53 pm